Download Positive Selection Driving the Evolution of a Gene of Male

Positive Selection Driving the Evolution of a Gene of Male Reproduction,
Acp26Aa, of Drosophila: II. Divergence Versus Polymorphism
Shun-Chern Tsaur, Chau-Ti Ting, and Chung-I Wu
Department of Ecology and Evolution, University of Chicago
The evolution of the gene for a male ejaculatory protein, Acp26Aa, has been shown to be driven by positive selection
when nonsibling species in the Drosophila melanogaster subgroup are compared. To know if selection has been
operating in the recent past and to understand the details of its dynamics, we obtained DNA sequences of Acp26Aa
and the nearby Acp26Ab gene from 39 D. melanogaster chromosomes. Together with the 10 published sequences,
we analyzed 49 sequences from five populations in four continents. The southern African population is somewhat
differentiated from all other populations, but its nucleotide diversity is lower at these two loci. We find the following
results for Acp26Aa: (1) The R : S (replacement : silent changes) ratio is significantly higher in the between-species
comparisons than in the within-species data by the McDonald and Kreitman test. Positive selection is probably
responsible for the excess of amino acid replacements between species. (2) However, within-species nucleotide
diversity is high. Neither the Tajima test nor the Fu and Li test indicates a reduction in nucleotide diversity due to
positive selection in the recent past. (3) The newly derived nucleotides in D. melanogaster are at high frequency
significantly more often than predicted by the neutral equilibrium. Since the nearby Acp26Ab gene does not show
these patterns, these observations cannot be attributed to the characteristics of this chromosomal region. We suggest
that positive selection is active, but may be weak, for each amino acid change in the Acp26Aa gene.
Introduction
The role sexual selection may play in the evolution
of hybrid male sterility has been suggested by a series
of studies on the genetic basis of speciation (Wu and
Davis 1993; Wu, Johnson, and Palopoli 1996; Hollocher
et al. 1997). A logical question to ask is whether the
signature of sexual selection can be observed at the molecular level. One might expect that genes specific for
male reproduction are more likely to be driven by sexual
selection and, hence, experience more rapid evolution
than genes of female or nonsexual functions. This logic
is best explained by Eberhard (1985) in his treatise on
the evolution of male genitalia.
The collection of accessory gland protein genes
(Acp) are excellent candidates for testing this hypothesis. This is a group of specialized proteins in the seminal
fluid of Drosophila (Chen 1984; Monsma and Wolfner
1988; Monsma, Harada, and Wolfner 1990; Park and
Wolfner 1995) that have been suggested to be variously
involved in egg-laying stimulation (Herndon and Wolfner 1995), remating inhibition (Chen 1976), female longevity (Fowler and Partridge 1989), and sperm competition (Clark et al. 1995), although the last aspect is not
conclusive (Herndon and Wolfner 1995). Among the
Acp genes, Acp26Aa and Acp26Ab have been most extensively analyzed for their evolutionary dynamics.
These two genes exist in tandem but show no sign of
homology (Monsma and Wolfner 1988). They are different in size (about 264 and 90 amino acids, respectively), and only the larger one, Acp26Aa, shows deviation from the neutral expectation in its DNA sequence
evolution (Aguade, Miyashita, and Langley 1992; Tsaur
and Wu 1997). We do not expect all male reproductive
Key words: Drosophila, accessory gland protein, sexual selection,
positive selection, population differentiation.
Address for correspondence and reprints: Chung-I Wu, Department of Ecology and Evolution, The University of Chicago, Chicago,
Illinois 60637. E-mail: [email protected].
Mol. Biol. Evol. 15(8):1040–1046. 1998
q 1998 by the Society for Molecular Biology and Evolution. ISSN: 0737-4038
1040
proteins to experience rapid evolution and, indeed, at
least two other Acp genes have been shown to evolve
at a moderate rate (Schmidt et al. 1993; Cirera and Aguade 1997). Our objective is to learn how and why some
of them, such as Acp26Aa, have evolved faster than the
neutral expectation (Kimura 1983).
The divergence of the Acp26Aa gene between distantly related species of Drosophila has been shown to
have the characteristics of genes driven by positive selection (Tsaur and Wu 1997). The most salient feature
is the higher rate (sometimes twice as high) of nonsynonymous than synonymous substitutions. In this study,
we ask whether this selective pressure has been a constant force driving the evolution of Acp26Aa or whether
it worked sporadically in the gene’s evolutionary past.
By examining the divergence of Acp26Aa between the
sibling species of Drosophila simulans and Drosophila
melanogaster, as well as the pattern of within-species
polymorphisms, we may gain insight into how often and
how strong positive selection has been at this locus. This
was also the focus of Aguade, Miyashita, and Langley
(1992), who sampled 10 sequences from a North American population. In this study, we add 39 sequences from
four other populations and have been able to perform
additional analyses.
Materials and Methods
Samples
We surveyed 39 isofemale lines of D. melanogaster
collected from four different continents: (1) 10 African
(Af) lines from Zimbabwe that have been described in
Begun and Aquadro (1993) and Hollocher et al. (1997);
(2) 10 Australian (Au) lines collected in 1995; (3) 11
Asian lines from Taiwan (Tw) collected in 1993 and
1996; and (4) 8 North American lines collected in upstate New York (Ny) in the late 1980s. We also include
the 10 published sequences (Aguade, Miyashita, and
Langley 1992) from flies collected in North Carolina
(Nc), bringing the total to 49 lines. To avoid heterozy-
Positive Selection on Acp26A
gosity, we used Df(2L)PM101, a deficiency that uncovers the region containing Acp26A, to extract single copies of the genes from each isofemale line. The DNA
sequences of this study are available from GenBank under accession numbers AF053250–AF053276 and
AF052470–AF052481.
PCR and DNA Sequencing
A 1.4-kb fragment encompassing both transcription
units of Acp26Aa and Acp26Ab was PCR-amplified by
the primers 1101 (59-ATGAACCAGATTTTATTATGC39) and 15492 (59-GCTTGAAGTAACTGCGG-39). Two
additional internal primers, 613 2 (59-TCCAAAAGTCTGCGGTTCAC-39) and 705 1 (59-CCCACCTGCCGCCAATC-39), were also used for sequencing. The 1 and 2 superscripts indicate the sense and
antisense strands, respectively. The numbers refer to the
59 nucleotide positions of the D. melanogaster sequence
as reported in Aguade, Miyashita, and Langley (1992).
Amplification was performed in 50 ml reaction volume.
Reactions were carried out for 30 cycles of 30 s at 948C
for denaturation, 30 s at 568C for annealing, 2 min at
728C for extension, and the final extension at 728C for
6 min. The desired fragment was then purified (Wizard
PCR Preps, Promega Corp., Madison, Wis.). Sequencing
was done on an ABI Prism 377 automated sequencer
using AmpliTaq DNA Polymerase, FS. All sequences
were determined from both strands.
Analysis
In the analysis of divergence between species, we
used only coding regions for the McDonald and Kreitman (1991) test. In this test, fixed differences are sites
where the nucleotide of the D. simulans sequence of
Aguade, Miyashita, and Langley (1992) is absent from
all 49 D. melanogaster sequences. These nucleotide
changes are then classified as either R (for amino acid
replacements) or S (for silent changes). In all other analyses (gene diversity, population differentation, frequency spectrum, and the HKA test), whole genes excluding
the primer sites were used. Acp26Aa corresponds to positions 47–898, and Acp26Ab corresponds to positions
1079–1413 of Aguade, Miyashita, and Langley (1992)
and figure 1. The sequences thus include intron, replacement, and silent sites.
We use H to designate nucleotide diversity and
present it as the average number of pairwise nucleotide
differences per kilobase. Following Hudson, Slatkin,
and Maddison (1992), Hw denotes the average nucleotide diversity within each population, and Hb denotes the
average nucleotide diversity between populations. Population differentiation, FST, is defined as 1 2 Hw/Hb.
Three statistical tests were performed on the entire
Acp26Aa and Acp26Ab genes: the HKA test (Hudson,
Kreitman, and Aguade 1987), the Tajima (1989) test,
and the Fu and Li (1993) test. These tests are explained
in the text. Calculations for nucleotide diversity, population differentiation, and the three tests were carried out
by using DnaSp (Rozas and Rozas 1997, version 2.51).
In the HKA test, we used the D. simulans sequence for
1041
the divergence estimate and pool all D. melanogaster
sequences for the polymorphism analysis.
Results
DNA sequences at polymorphic sites are presented
in figure 1. In total, there are 60 segregating sites in the
sample of 49 sequences from five populations.
Divergence Between Species
The first question we ask is whether the amino acid
substitutions between the Acp26A genes of D. melanogaster and D. simulans can be accounted for by neutral
evolution. Since the rate of nonsynonymous substitutions is not significantly different from that of synonymous substitutions in Acp26Aa (Aguade, Miyashita, and
Langley 1992), we do not know whether the protein
experiences no selection or whether positive selection,
which accelerates substitutions, is roughly cancelled out
by negative selection, which retards them. To address
this issue, we compared the level of polymorphism with
that of divergence for both synonymous changes (S) and
amino acid replacements (R), a procedure pioneered by
McDonald and Kreitman (1991). The results are presented in table 1.
Under the neutral theory, the R : S ratio for fixed
differences between species should be the same as that
for within-species polymorphisms. The ratios for
Acp26Aa, 3.57 and 1.38, respectively, are significantly
different at the 5% level. On the other hand, the ratios
for Acp26Ab are comparable (although the level of divergence is too small to be truly informative). Since the
ratio of the number of replacement sites over that of
silent sites is about 3:1, the low R : S value for the
Acp26Aa polymorphism suggests that negative selection
against amino acid changes has been operating. The near
parity in the rates of synonymous and replacement substitutions between species thus cannot be explained by
neutral evolution over the entire gene. The number of
amino acid replacements between species must have
been enhanced by positive selection.
The higher-than-expected level of divergence in the
Acp26Aa gene can also be seen in the comparison with
the Acp26Ab gene by means of the HKA test. The test
compares the level of within-species polymorphism with
that of between-species divergence for two different loci
(see Analysis in Materials and Methods). It yields a significant x2 value (7.27, df 5 1, P , 0.01), indicating
that the divergence/polymorphism ratio is higher in the
Acp26Aa gene than in Acp26Ab. Much of the difference
is due to the high level of divergence in the former.
Variation Within Species
To explain the excess in the number of amino acid
substitutions between D. simulans and D. melanogaster,
the most obvious hypothesis is directional selection, under which a favorable mutation increases its frequency
until fixation. These selection-driven events themselves
contribute very little to the observed level of withinspecies variation, because they introduce only transient
polymorphism into the population. An important consequence of positive selection is that, if it has happened
1042
Tsaur et al.
FIG. 1.—Nucleotides at polymorphic sites of the Acp26Aa and Acp26Ab genes. The numbering of nuclotide position follows that of Aguade,
Miyashita, and Langley (1992). The gene structure in relation to the nuclotide position is given in figure 3. A dot denotes identity with the Af1
sequence. Each sequence label denotes the geographical origin (see Materials and Methods).
in the recent past, a reduction in the level of polymorphism into the nearby region is expected (Maynard
Smith and Haigh 1974; Kaplan, Hudson, and Langley
1989; Braverman et al. 1995; Fu 1996). This process is
sometimes referred to as a ‘‘selective sweep.’’ A rough
estimation based on table 1 suggests an excess of 46
(575 2 [22/16] 3 21) amino acid replacements between D. melanogaster and D. simulans that may be
attributable to positive selection. (This is because, under
the neutral model, a parity in R : S values between polymorphic and fixed differences is expected). If these replacements occurred independently in the evolutionary
past, two or three may have been sufficiently recent to
have an impact on the level of polymorphism. We wish
to know if we can detect such a footprint of positive
selection within D. melanogaster.
Since the sequences originate from five different
populations, we first analyze the levels of nucleotide diversity within and between populations as shown in table 2. Although the level of nucleotide diversity is somewhat lower in Acp26Aa than in Acp26Ab, neither exhibits a reduction in polymorphism in comparison with other genes (Moriyama and Powell 1996). Table 2 also
confirms that the Zimbabwe population of Africa is dif-
Positive Selection on Acp26A
Table 1
Nucleotide Polymorphism Within Drosophila melanogaster
and Divergence Between D. melanogaster and Drosophila
simulans for Acp26Aa and Acp26Ab
POLYMORPHIC
FIXED DIFFERENCES
(between species)
Acp26Aa . . .
Acp26Ab . . .
DIFFERENCES
(within species)
R
S
R
S
G VALUE
75
1
21
2
22
4
16
6
5.33*
0.044
NOTE.—R and S stand for amino acid replacements and silent changes, respectively.
* P , 0.05.
ferentiated from populations of other continents at the
molecular level (Begun and Aquadro 1993). However,
part of the differentiation is due to the reduced nucleotide diversity of the Zimbabwe population, an observation that is contrary to the pattern reported for X-linked
genes (Begun and Aquadro 1993). In the rest of this
study, we shall concentrate on the Acp26Aa gene, which
appears to be influenced by positive selection (Aguade,
Miyashita, and Langley 1992; Tsaur and Wu 1997).
Since Acp26Aa shows little differentiation among the
four non-African populations and only moderate differentiation between the African and non-African populations, we shall combine the data from all populations.
A normal level of nucleotide diversity does not
necessarily mean the absence of a recent selective
sweep. If the region evolves faster than other regions
due to a greater mutation rate, for example, the population can recover quickly from the loss of diversity
swept away by positive selection. (The synonymous
substitution rate in this region is indeed higher than are
those in other regions [Aguade, Miyashita, and Langley
1992; Tsaur and Wu 1997].) One characteristic of populations in the process of recovering nucleotide diversity
lost through selective sweep or population bottleneck is
the excess of rare or recent mutations relative to the total
amount of variation observed (Nei, Maruyama, and
Chakraborty 1977). In Tajima’s (1989) test and Fu and
Li’s (1993) test, a negative statistic indicates an excess
of rare or recent mutations. Both tests for both genes
yield negative values, but none is significant at even the
10% level (table 3). Therefore, the overall nucleotide
diversity and the test statistics related to it fail to yield
evidence of a recent selective sweep.
In general, the tests above utilize only a portion of
the information on the mutant frequency. For example,
Fu and Li’s (1993) test examines whether the relative
abundance of singletons in the sample (variants that occur only once) agrees with the neutral prediction. In figure 2, the entire frequency spectrum of Acp26Aa is displayed. On the abscissa is the number of occurrences of
the new nucleotide, which ranges from 1 to 48 in the
sample of 49. On the ordinate is the number of sites
having a particular occurrence. It is possible to determine the new nucleotide in D. melanogaster because the
sequences from its three sibling species are available
(Aguade, Miyashita, and Langley 1992). We only con-
1043
Table 2
Between- and Within-Population Nucleotide Diversity and
the FST Measure of Population Differentiation
Af
Au
Nc
Acp26Aa
Af . . . . . .
Au . . . . .
Nc . . . . .
Ny . . . . .
Tw . . . . .
Ny
Tw
5.7
0.219
0.133
0.183
0.081
9.3
9.0
0.023
20.030
0.113
7.7
8.4
7.7
20.014
0.037
8.7
8.4
7.9
8.6
0.023
6.9
8.9
7.6
7.9
7.0
Acp26Ab
Af . . . . . .
Au . . . . .
Nc . . . . .
Ny . . . . .
Tw . . . . .
4.9
0.398
0.392
0.270
0.207
12.2
9.9
0.013
0.081
0.071
12.0
9.9
9.9
0.001
20.038
11.4
11.8
10.8
11.9
20.035
12.6
13.5
12.0
13.0
15.1
NOTE.—For population designation and statistical definition, see Materials
and Methods. Above diagonals, Hb (between-population nucleotide diversity);
diagonals, Hw (within-population nucleotide diversity); below diagonals, FST 5
1 2 Hw/Hb (Hudson, Slatkin, and Maddison 1992).
sider polymorphic sites that are unambiguous with respect to the ancestral versus derived nucleotide. There
are 31 such sites. (The ancestral state is determined to
be the particular nucleotide of the polymorphism which
is also present in all three sibling species.)
At the neutral equilibrium, the expected number of
sites at which the new nucleotide is present i times in
the sample is given by 4Nv/i (Fu 1995), where N and v
are the effective population size and mutation rate, respectively. To compute the expected spectrum, the estimated 4Nv according to Watterson’s (1975) formula is
the number of observed sites (31) divided by (1 1 ½ 1
⅓ 1 . . . 1 1⁄48). The expected and observed frequency
spectra are both given in figure 2. There appear to be
too many sites with the new nucleotide in high frequency (.0.5): 10 of the 31 are such sites. The expected
values for the high- and low-frequency mutations are
5.04 and 25.96, respectively. A x2 test between the observed and expected numbers yields a value of 5.83,
which is significant at the 5% level (df 5 1).
The presence of the many sites with the derived
nucleotide at a high frequency also explains the discrepancy in the test statistics between Tajima’s and Fu
and Li’s tests shown in table 3 for Acp26Aa (20.875 vs.
20.118). This is because whether a new mutation is at
a frequency of x or 1 2 x would not have an effect on
Tajima’s test, whereas the two situations may yield very
different results in Fu and Li’s test. Tajima’s test examines the relative frequency of the rare alleles, while
Table 3
Summary Statistics of the Polymorphic Data
na
sb
Hc
Tajima’s D
Fu and Li’s
D
Acp26Aa . . 49
Acp26Ab . . 49
37
18
8.01 6 0.57
11.67 6 0.79
20.875d
20.109d
20.118d
21.294d
a
Number of sequences.
Number of polymorphic sites.
Average nucleotide diversity (bp differences per kb).
d No D value is significant.
b
c
1044
Tsaur et al.
FIG. 2.—The observed and expected frequency spectra of the newly derived nucleotide in Acp26Aa of D. melanogaster. The abscissa is
the number of occurrences (i) of the new nucleotide, ranging from 1 to 48 in a sample of 49 sequences. The ordinate is the number of sites
with i occurrences. There are 31 sites where the derived nucleotide can be unambiguously inferred with reference to the three outgroup species.
The expected frequency spectrum is given by Watterson (1975; see text). We divide the occurrences into high ($24) and low (,24) frequency
classes for the goodness-of-fit test.
Fu and Li’s test studies the frequency of the young alleles. The two tests should be comparable except in
dealing with unusual frequency spectra like that of figure 2. The difference in these two tests for Acp26Ab
may also be due to an unusual frequency spectrum, although the number of polymorphic sites is too small to
be statistically trackable.
A more appropriate statistic than that provided by
the goodness-of-fit test presented above may be difficult.
Fu (1996) gives perhaps the most rigorous test for deviation of an observed frequency spectrum from the
neutral expectation but his test generally requires the
absence of recombination. As shown in figure 3, there
is no indication of extensive linkage disequilibrium, and
in a separate analysis, recombinations were often detectable even between sites less than 50 bp apart.
Discussion
The most parsimonious interpretation of the results
presented in table 1, in conjunction with the observations of Tsaur and Wu (1997), is that positive selection
drives the divergence between species at the Acp26Aa
locus. This selection has been operating in both the distant and recent evolutionary past of the D. melanogaster
species subgroup. However, if the excess of amino acid
substitutions between D. melanogaster and D. simulans
is due to positive selection, there is no sign of a recent
selective sweep resulting in the loss of variation within
the species of D. melanogaster. Because our estimate of
the number of selectively driven events between the two
species is necessarily imprecise, it is possible that the
actual number is much smaller and the last sweep occurred too long ago to leave a trace.
Alternatively, positive selection may have been
constantly at work but too weak to depress the nucleotide diversity, thus leaving no trace detectable by either
Tajima’s (1989) or Fu and Li’s (1993) test. Weak selection would take a long time to drive a favorable mutation to fixation, and therefore even very closely linked
regions may escape the sweep through recombination.
On the other hand, this selection should still be orders
of magnitude stronger than genetic drift, resulting in the
acceleration of between-species divergence. Is there a
range of values for the selective coefficient that satisfy
both conditions? Based on the results of Kaplan, Hudson, and Langley (1989) and the parameters from Drosophila, even sites as close as 50 bp may not share an
identical history if 2Ns is less than 103 (s being the
selective advantage at one of the two sites). Since 2Ns
. 10 is stronger than genetic drift, there exist at least
two orders of magnitude for which selection is strong
enough to overpower genetic drift but still weak enough
to have only a restricted footprint (,50 bp) on nucleotide diversity (see Braverman et al. 1995 for a detailed
study). Such weak hitchhiking can be more effectively
detected by examining the frequency spectrum of new
mutations, as shown in figure 2. Diversity reduction may
be viewed as an extreme example of strong selection
Positive Selection on Acp26A
1045
FIG. 3.—Linkage disequilibria in all pairwise comparisons among the 60 segregating sites in the Acp26A region. Data are based on all 49
sequences. The locations of the segregating sites relative to the two structural genes are shown on the diagonal. The levels of significance by
Fisher’s exact test are shown by different shades. Black, P , 0.001; dark gray, 0.001 , P , 0.01; light gray, 0.01 , P , 0.05.
disrupting gene frequency spectrum, while weaker selection may leave a different signature by dragging nearby variations to higher-than-expected frequencies without causing fixation.
In inferring the derived nucleotide in figure 2, we
use the sequences from the three sibling species as outgroups and apply the parsimony principle. The question
is how well parsimony holds in this data set. We shall
compare the two types of nucleotide sites, (C ↔ T) 5
[T, T, T] and (C ↔ T) ↔ [A, A, A], where ( ) contains
the polymorphic nucleotides of D. melanogaster, [ ]
contains the nucleotides of the three sibling species, 5
denotes no change, and ↔ denotes a nucleotide change.
(The nucleotides are chosen merely for demonstration.)
The former set is usually inferred to have experienced
a T-to-C change, as shown. If parsimony is violated,
then C may in fact be the ancestral type which changed
to T twice during evolution, i.e., (T ↔ C) ↔ [T, T, T].
Given the same number of changes, this set should be
half as frequent as (T ↔ C) ↔ [R, R, R], where R is
either A or G. Since we observed no site with this latter
type of nucleotide set, the parsimony inference seems
adequate in our analysis.
In performing the statistical tests, we combine the
data from five populations. This should not affect our
interpretation. For the D statistics, Tajima (1993) has
shown that the degree of population differentiation observed in our sample would not affect the results much
at all. Similarly, Fu’s studies (1996, see his fig. 1) suggest that the entire frequency spectrum is not greatly
affected by population structure or sampling scheme.
In conclusion, positive selection on the amino acid
sequence of Acp26Aa may have been a constant feature
during the evolutionary history of the D. melanogaster
subgroup. The selective coefficient for each amino acid
substitution is likely to be small, at most two to three
orders of magnitude stronger than the effect of genetic
drift.
Acknowledgments
The authors are indebted to Ian Boussy and YengYu Yang for flies from Australia and Taiwan, Yun-Xin
Fu for statistical consultations, and Mariana Wolfner for
discussions. We also thank Casey Bergman, Mark Jensen, and Jerry Wyckoff for comments on the manuscript.
This work was supported by NSF and NIH grants to
C.-I W.
LITERATURE CITED
AGUADE, M., N. MIYASHITA, and C. H. LANGLEY. 1992. Polymorphism and divergence in the Mst26A male accessory
gland gene region in Drosophila. Genetics 132:755–770.
BRAVERMAN, J. M., R. R. HUDSON, N. L. KAPLAN, C. H.
LANGLEY, and W. STEPHAN. 1995. The hitchhiking effect
on the site frequency spectrum of DNA polymorphisms.
Genetics 140:783–796.
BEGUN, D. J., and C. F. AQUADRO. 1993. African and North
American populations of Drosophila melanogaster are very
different at the DNA level. Nature 356:519–520.
CHEN, P. S. 1976. Species-specific protein patterns in Drosophila accessory glands. Experientia 32:549–550.
. 1984. The functional morphology and biochemistry of
insect male accessory glands and their secretion. Annu.
Rev. Entomol. 29:233–255.
CIRERA, S., and M. AGUADE. 1997. Evolutionary history of the
sex-peptide (Acp70A) gene region in Drosophila melanogaster. Genetics 147:189–197.
1046
Tsaur et al.
CLARK, A. G., M. AGUADE, T. PROUT, L. G. HARSHMAN, and
C. H. LANGLEY. 1995. Variation in sperm displacement and
its association with accessory gland protein loci in Drosophila melanogaster. Genetics 139:189–201.
EBERHARD, W. G. 1985. Sexual selection and animal genetalia.
Harvard University Press, Cambridge, Mass.
FOWLER, K., and L. PARTRIDGE. 1989. A cost of mating in
female fruitflies. Nature 338:760–761.
FU, Y. X. 1995. Statistical properties of segregating sites.
Theor. Popul. Biol. 48:172–197.
1996. New statistical tests of neutrality for DNA samples from a population. Genetics 143:557–570.
FU, Y. X., and W.-H. LI. 1993. Statistical tests of neutrality of
mutations. Genetics 133:693–709.
HERNDON, L. A., and M. F. WOLFNER. 1995. A Drosophila
seminal fluid protein, Acp26Aa, stimulates egg laying in
females for 1 day after mating. Proc. Natl. Acad. Sci. USA
92:10114–10118.
HOLLOCHER, H., C. T. TING, F. POLLACK, and C.-I WU. 1997.
Incipient speciation by sexual isolation in Drosophila melanogaster: variation in mating preference and correlation
between sexes. Evolution 51:1175–1181.
HUDSON, R. R., M. KREITMAN, and M. AGUADE. 1987. A test
of neutral molecular evolution based on nucleotide data.
Genetics 116:153–159.
HUDSON, R. R., M. SLATKIN, and M. P. MADDISON. 1992. Estimation of level of gene flow from DNA sequence data.
Genetics 132:583–589.
KAPLAN, N. L., R. R. HUDSON, and C. H. LANGLEY. 1989. The
‘‘hitchhiking effect’’ revisited. Genetics 123:887–899.
KIMURA, M. 1983. The neutral theory of molecular evolution.
Cambridge University Press, Cambridge, UK.
MCDONALD, J. H., and M. KREITMAN. 1991. Adaptative protein evolution at the Adh locus in Drosophila. Nature 351:
652–654.
MAYNARD SMITH, J., and J. HAIGH. 1974. The hitch-hiking
effect of a favorable gene. Genet. Res. 23:23–35.
MONSMA, S. A., H. A. HARADA, and M. F. WOLFNER. 1990.
Synthesis of two Drosophila male accessory gland proteins
and their fate after transfer to the female during mating.
Dev. Biol. 142:465–475.
MONSMA, S. A., and M. F. WOLFNER. 1988. Structure and expression of a Drosophila male accessory gland gene whose
product resembles a peptide pheromone precursor. Genes
Dev. 2:1063–1073.
MORIYAMA, E. N., and J. R. POWELL. 1996. Intraspecific nuclear DNA variation in Drosophila. Mol. Biol. Evol. 13:
261–277.
NEI, M., T. MARUYAMA, and R. CHAKRABORTY. 1975. The bottleneck effect and genetic variability in populations. Evolution 29:1–10.
PARK, M., and M. F. WOLFNER. 1995. Male and female cooperate in the pheromone-like processing of a Drosophila
melanogaster seminal fluid protein. Dev. Biol. 171:694–
702.
ROZAS, J., and R. ROZAS. 1997. DnaSp version 2.0: a novel
software package for extensive molecular population genetics analysis. Comput. Appl. Biosci. 13:307–311.
SCHMIDT, T., Y. CHOFFAT, M. SCHNEIDER, P. HUNZIKER, Y.
FUYAMA, and E. KUBLI. 1993. Drosophila suzukii contains
a peptide homologous to the Drosophila melanogaster sex
peptide and functional in both species. Insect Biochem.
Mol. Biol. 23:571–579.
TAJIMA, F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123:
585–595.
. 1993. Measurement of DNA polymorphism. Pp. 37–
59 in N. TAKAHATA and A. G. CLARK, eds. Mechanisms of
molecular evolution. Sinauer, Sunderland, Mass.
TSAUR, S. C., and C.-I WU. 1997. Positive selection and the
molecular evolution of a gene of male reproduction,
Acp26Aa of Drosophila. Mol. Biol. Evol. 14:544–549.
WATTERSON, G. A. 1975. On the number of segregating sites.
Theor. Popul. Biol. 7:256–276.
WU, C.-I, and A. W. DAVIS. 1993. Evolution of postmating
reproductive isolation: the composite nature of Haldane’s
rule and its genetic bases. Am. Nat. 142:187–212.
WU, C.-I, N. A. JOHNSON, and M. F. PALOPOLI. 1996. Haldane’s
rule and its legacy: why are there so many sterile males?
Trends Ecol. Evol. 11:281–284.
CHARLES F. AQUADRO, reviewing editor
Accepted April 27, 1998